Merge branch 'main' into kevin

.github/pull_request_template.md

@@ -1,8 +1,6 @@
-- [ ] Include this change in the Release Notes. If checked, you must provide an **end-user friendly** description for your change below
-
 **End-user friendly description of the problem this fixes or functionality that this introduces**
 
-
+- [ ] Include this change in the Release Notes. If checked, you must provide an **end-user friendly** description for your change below
 
 ---
 **Give a summary of what the PR does, explaining any non-trivial design decisions**

.github/workflows/dummy-agent-test.yml

@@ -45,7 +45,7 @@ jobs:
       - name: Run tests
         run: |
           set -e
-          poetry run python3 openhands/core/main.py -t "do a flip" -d ./workspace/ -c DummyAgent
+          SANDBOX_FORCE_REBUILD_RUNTIME=True poetry run python3 openhands/core/main.py -t "do a flip" -d ./workspace/ -c DummyAgent
       - name: Check exit code
         run: |
           if [ $? -ne 0 ]; then

.github/workflows/ghcr-build.yml

@@ -293,7 +293,7 @@ jobs:
           SANDBOX_RUNTIME_CONTAINER_IMAGE=$image_name \
           TEST_IN_CI=true \
           RUN_AS_OPENHANDS=false \
-          poetry run pytest -n 3 -raR --reruns 1 --reruns-delay 3 --cov=agenthub --cov=openhands --cov-report=xml -s ./tests/runtime
+          poetry run pytest -n 3 -raRs --reruns 2 --reruns-delay 5 --cov=agenthub --cov=openhands --cov-report=xml -s ./tests/runtime
       - name: Upload coverage to Codecov
         uses: codecov/codecov-action@v4
         env:
@@ -371,7 +371,7 @@ jobs:
           SANDBOX_RUNTIME_CONTAINER_IMAGE=$image_name \
           TEST_IN_CI=true \
           RUN_AS_OPENHANDS=true \
-          poetry run pytest -n 3 -raR --reruns 1 --reruns-delay 3 --cov=agenthub --cov=openhands --cov-report=xml -s ./tests/runtime
+          poetry run pytest -n 3 -raRs --reruns 2 --reruns-delay 5 --cov=agenthub --cov=openhands --cov-report=xml -s ./tests/runtime
       - name: Upload coverage to Codecov
         uses: codecov/codecov-action@v4
         env:

openhands/core/cli.py

@@ -91,7 +91,7 @@ async def main():
         return
 
     logger.setLevel(logging.WARNING)
-    config = load_app_config()
+    config = load_app_config(config_file=args.config_file)
     sid = 'cli'
 
     agent_cls: Type[Agent] = Agent.get_cls(config.default_agent)

openhands/core/config/sandbox_config.py

@@ -18,6 +18,7 @@ class SandboxConfig:
         enable_auto_lint: Whether to enable auto-lint.
         use_host_network: Whether to use the host network.
         initialize_plugins: Whether to initialize plugins.
+        force_rebuild_runtime: Whether to force rebuild the runtime image.
         runtime_extra_deps: The extra dependencies to install in the runtime image (typically used for evaluation).
             This will be rendered into the end of the Dockerfile that builds the runtime image.
             It can contain any valid shell commands (e.g., pip install numpy).
@@ -46,6 +47,7 @@ class SandboxConfig:
     )
     use_host_network: bool = False
     initialize_plugins: bool = True
+    force_rebuild_runtime: bool = False
     runtime_extra_deps: str | None = None
     runtime_startup_env_vars: dict[str, str] = field(default_factory=dict)
     browsergym_eval_env: str | None = None

openhands/core/config/utils.py

@@ -289,6 +289,12 @@ def get_llm_config_arg(
 def get_parser() -> argparse.ArgumentParser:
     """Get the parser for the command line arguments."""
     parser = argparse.ArgumentParser(description='Run an agent with a specific task')
+    parser.add_argument(
+        '--config-file',
+        type=str,
+        default='config.toml',
+        help='Path to the config file (default: config.toml in the current directory)',
+    )
     parser.add_argument(
         '-d',
         '--directory',
@@ -383,14 +389,17 @@ def parse_arguments() -> argparse.Namespace:
     return parsed_args
 
 
-def load_app_config(set_logging_levels: bool = True) -> AppConfig:
-    """Load the configuration from the config.toml file and environment variables.
+def load_app_config(
+    set_logging_levels: bool = True, config_file: str = 'config.toml'
+) -> AppConfig:
+    """Load the configuration from the specified config file and environment variables.
 
     Args:
         set_logger_levels: Whether to set the global variables for logging levels.
+        config_file: Path to the config file. Defaults to 'config.toml' in the current directory.
     """
     config = AppConfig()
-    load_from_toml(config)
+    load_from_toml(config, config_file)
     load_from_env(config, os.environ)
     finalize_config(config)
     if set_logging_levels:

openhands/core/main.py

@@ -228,7 +228,7 @@ def generate_sid(config: AppConfig, session_name: str | None = None) -> str:
     # Load the app config
     # this will load config from config.toml in the current directory
     # as well as from the environment variables
-    config = load_app_config()
+    config = load_app_config(config_file=args.config_file)
 
     # Override default LLM configs ([llm] section in config.toml)
     if args.llm_config:

openhands/llm/llm.py

@@ -18,6 +18,7 @@
 from litellm.caching import Cache
 from litellm.exceptions import (
     APIConnectionError,
+    APIError,
     InternalServerError,
     RateLimitError,
     ServiceUnavailableError,
@@ -39,6 +40,7 @@
 # tuple of exceptions to retry on
 LLM_RETRY_EXCEPTIONS: tuple[type[Exception], ...] = (
     APIConnectionError,
+    APIError,
     InternalServerError,
     RateLimitError,
     ServiceUnavailableError,

openhands/runtime/builder/docker.py

@@ -113,8 +113,8 @@ def build(
                 raise subprocess.CalledProcessError(
                     return_code,
                     process.args,
-                    output=None,
-                    stderr=None,
+                    output=process.stdout.read() if process.stdout else None,
+                    stderr=process.stderr.read() if process.stderr else None,
                 )
 
         except subprocess.CalledProcessError as e:

openhands/runtime/builder/remote.py

@@ -9,7 +9,7 @@
 from openhands.runtime.builder import RuntimeBuilder
 from openhands.runtime.utils.request import send_request
 from openhands.runtime.utils.shutdown_listener import (
-    should_exit,
+    should_continue,
     sleep_if_should_continue,
 )
 
@@ -60,8 +60,8 @@ def build(self, path: str, tags: list[str]) -> str:
         # Poll /build_status until the build is complete
         start_time = time.time()
         timeout = 30 * 60  # 20 minutes in seconds
-        while True:
-            if should_exit() or time.time() - start_time > timeout:
+        while should_continue():
+            if time.time() - start_time > timeout:
                 logger.error('Build timed out after 30 minutes')
                 raise RuntimeError('Build timed out after 30 minutes')
 
@@ -101,6 +101,8 @@ def build(self, path: str, tags: list[str]) -> str:
             # Wait before polling again
             sleep_if_should_continue(30)
 
+        raise RuntimeError('Build interrupted (likely received SIGTERM or SIGINT).')
+
     def image_exists(self, image_name: str, pull_from_repo: bool = True) -> bool:
         """Checks if an image exists in the remote registry using the /image_exists endpoint."""
         params = {'image': image_name}

openhands/runtime/client/runtime.py

@@ -177,11 +177,13 @@ def __init__(
                     raise ValueError(
                         'Neither runtime container image nor base container image is set'
                     )
+                logger.info('Preparing container, this might take a few minutes...')
                 self.send_status_message('STATUS$STARTING_CONTAINER')
                 self.runtime_container_image = build_runtime_image(
                     self.base_container_image,
                     self.runtime_builder,
                     extra_deps=self.config.sandbox.runtime_extra_deps,
+                    force_rebuild=self.config.sandbox.force_rebuild_runtime,
                 )
             self.container = self._init_container(
                 sandbox_workspace_dir=self.config.workspace_mount_path_in_sandbox,
@@ -302,7 +304,7 @@ def _init_container(
             container = self.docker_client.containers.run(
                 self.runtime_container_image,
                 command=(
-                    f'/openhands/miniforge3/bin/mamba run --no-capture-output -n base '
+                    f'/openhands/micromamba/bin/micromamba run -n openhands '
                     f'poetry run '
                     f'python -u -m openhands.runtime.client.client {self._container_port} '
                     f'--working-dir "{sandbox_workspace_dir}" '
@@ -450,13 +452,15 @@ def run_action(self, action: Action) -> Observation:
                     logger.debug(f'response: {response}')
                     error_message = response.text
                     logger.error(f'Error from server: {error_message}')
-                    obs = ErrorObservation(f'Command execution failed: {error_message}')
+                    obs = ErrorObservation(f'Action execution failed: {error_message}')
             except requests.Timeout:
                 logger.error('No response received within the timeout period.')
-                obs = ErrorObservation('Command execution timed out')
+                obs = ErrorObservation(
+                    f'Action execution timed out after {action.timeout} seconds.'
+                )
             except Exception as e:
-                logger.error(f'Error during command execution: {e}')
-                obs = ErrorObservation(f'Command execution failed: {str(e)}')
+                logger.error(f'Error during action execution: {e}')
+                obs = ErrorObservation(f'Action execution failed: {str(e)}')
             self._refresh_logs()
             return obs
 

openhands/runtime/plugins/jupyter/__init__.py

@@ -28,7 +28,8 @@ async def initialize(self, username: str, kernel_id: str = 'openhands-default'):
                 'cd /openhands/code\n'
                 'export POETRY_VIRTUALENVS_PATH=/openhands/poetry;\n'
                 'export PYTHONPATH=/openhands/code:$PYTHONPATH;\n'
-                '/openhands/miniforge3/bin/mamba run -n base '
+                'export MAMBA_ROOT_PREFIX=/openhands/micromamba;\n'
+                '/openhands/micromamba/bin/micromamba run -n openhands '
                 'poetry run jupyter kernelgateway '
                 '--KernelGatewayApp.ip=0.0.0.0 '
                 f'--KernelGatewayApp.port={self.kernel_gateway_port}\n'

openhands/runtime/remote/runtime.py

@@ -119,6 +119,7 @@ def __init__(
                 self.config.sandbox.base_container_image,
                 self.runtime_builder,
                 extra_deps=self.config.sandbox.runtime_extra_deps,
+                force_rebuild=self.config.sandbox.force_rebuild_runtime,
             )
 
             response = send_request(
@@ -144,8 +145,8 @@ def __init__(
         start_request = {
             'image': self.container_image,
             'command': (
-                f'/openhands/miniforge3/bin/mamba run --no-capture-output -n base '
-                'PYTHONUNBUFFERED=1 poetry run '
+                f'/openhands/micromamba/bin/micromamba run -n openhands '
+                'poetry run '
                 f'python -u -m openhands.runtime.client.client {self.port} '
                 f'--working-dir {self.config.workspace_mount_path_in_sandbox} '
                 f'{plugin_arg}'

openhands/runtime/utils/runtime_templates/Dockerfile.j2

@@ -1,11 +1,13 @@
-{% if skip_init %}
 FROM {{ base_image }}
-{% else %}
+
+# Shared environment variables (regardless of init or not)
+ENV POETRY_VIRTUALENVS_PATH=/openhands/poetry
+ENV MAMBA_ROOT_PREFIX=/openhands/micromamba
+
+{% if not skip_init %}
 # ================================================================
 # START: Build Runtime Image from Scratch
 # ================================================================
-FROM {{ base_image }}
-
 {% if 'ubuntu' in base_image and (base_image.endswith(':latest') or base_image.endswith(':24.04')) %}
 {% set LIBGL_MESA = 'libgl1' %}
 {% else %}
@@ -14,7 +16,7 @@ FROM {{ base_image }}
 
 # Install necessary packages and clean up in one layer
 RUN apt-get update && \
-    apt-get install -y wget sudo apt-utils {{ LIBGL_MESA }} libasound2-plugins git docker.io lsof && \
+    apt-get install -y wget curl sudo apt-utils {{ LIBGL_MESA }} libasound2-plugins git docker.io lsof && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
@@ -26,19 +28,16 @@ RUN mkdir -p /openhands && \
     mkdir -p /openhands/logs && \
     mkdir -p /openhands/poetry
 
-# Directory containing subdirectories for virtual environment.
-ENV POETRY_VIRTUALENVS_PATH=/openhands/poetry
+# Install micromamba
+RUN mkdir -p /openhands/micromamba/bin && \
+    /bin/bash -c "PREFIX_LOCATION=/openhands/micromamba BIN_FOLDER=/openhands/micromamba/bin INIT_YES=no CONDA_FORGE_YES=yes $(curl -L https://micro.mamba.pm/install.sh)" && \
+    /openhands/micromamba/bin/micromamba config remove channels defaults && \
+    /openhands/micromamba/bin/micromamba config list
 
-RUN if [ ! -d /openhands/miniforge3 ]; then \
-    wget --progress=bar:force -O Miniforge3.sh "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh" && \
-    bash Miniforge3.sh -b -p /openhands/miniforge3 && \
-    rm Miniforge3.sh && \
-    chmod -R g+w /openhands/miniforge3 && \
-    bash -c ". /openhands/miniforge3/etc/profile.d/conda.sh && conda config --set changeps1 False && conda config --append channels conda-forge"; \
-    fi
+# Create the openhands virtual environment and install poetry and python
+RUN /openhands/micromamba/bin/micromamba create -n openhands -y && \
+    /openhands/micromamba/bin/micromamba install -n openhands -c conda-forge poetry python=3.12 -y
 
-# Install Python and Poetry
-RUN /openhands/miniforge3/bin/mamba install conda-forge::poetry python=3.12 -y
 # ================================================================
 # END: Build Runtime Image from Scratch
 # ================================================================
@@ -59,27 +58,28 @@ COPY ./code /openhands/code
 # virtual environment are used by default.
 WORKDIR /openhands/code
 RUN \
+    /openhands/micromamba/bin/micromamba config set changeps1 False && \
     # Configure Poetry and create virtual environment
-    /openhands/miniforge3/bin/mamba run -n base poetry config virtualenvs.path /openhands/poetry && \
-    /openhands/miniforge3/bin/mamba run -n base poetry env use python3.11 && \
+    /openhands/micromamba/bin/micromamba run -n openhands poetry config virtualenvs.path /openhands/poetry && \
+    /openhands/micromamba/bin/micromamba run -n openhands poetry env use python3.11 && \
     # Install project dependencies
-    /openhands/miniforge3/bin/mamba run -n base poetry install --only main,runtime --no-interaction --no-root && \
+    /openhands/micromamba/bin/micromamba run -n openhands poetry install --only main,runtime --no-interaction --no-root && \
     # Update and install additional tools
     apt-get update && \
-    /openhands/miniforge3/bin/mamba run -n base poetry run pip install playwright docker && \
-    /openhands/miniforge3/bin/mamba run -n base poetry run playwright install --with-deps chromium && \
+    /openhands/micromamba/bin/micromamba run -n openhands poetry run pip install playwright docker&& \
+    /openhands/micromamba/bin/micromamba run -n openhands poetry run playwright install --with-deps chromium && \
     # Set environment variables
-    echo "OH_INTERPRETER_PATH=$(/openhands/miniforge3/bin/mamba run -n base poetry run python -c "import sys; print(sys.executable)")" >> /etc/environment && \
+    echo "OH_INTERPRETER_PATH=$(/openhands/micromamba/bin/micromamba run -n openhands poetry run python -c "import sys; print(sys.executable)")" >> /etc/environment && \
     # Install extra dependencies if specified
     {{ extra_deps }} {% if extra_deps %} && {% endif %} \
     # Clear caches
-    /openhands/miniforge3/bin/mamba run -n base poetry cache clear --all . && \
+    /openhands/micromamba/bin/micromamba run -n openhands poetry cache clear --all . && \
     # Set permissions
     {% if not skip_init %}chmod -R g+rws /openhands/poetry && {% endif %} \
     mkdir -p /openhands/workspace && chmod -R g+rws,o+rw /openhands/workspace && \
     # Clean up
     apt-get clean && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
-    /openhands/miniforge3/bin/mamba clean --all
+    /openhands/micromamba/bin/micromamba clean --all
 # ================================================================
 # END: Copy Project and Install/Update Dependencies
 # ================================================================

openhands/server/listen.py

@@ -1,3 +1,4 @@
+import asyncio
 import os
 import re
 import tempfile
@@ -257,7 +258,7 @@ async def websocket_endpoint(websocket: WebSocket):
         {"action": "finish", "args": {}}
         ```
     """
-    await websocket.accept()
+    await asyncio.wait_for(websocket.accept(), 10)
 
     if websocket.query_params.get('token'):
         token = websocket.query_params.get('token')